Views and News 
Hi everyone.
I see that this thread which I started on the 8th of October has passed the 28000 views and is now 6th out of 5890 threads for number of views in the solid state amplifier section (and seventh for number of replies).
It is a matter of some surprise to me the interest and level of discussion that has occurred and it is very pleasing to hear of people's progress with their versions of my design.
I do my best to answer questions as fully as possible but I am aware that in all the replies there must be some which I have inadvertently overlooked.
As it is the last few weeks I have probably been spending up to 50% of my available "work" time directly or indirectly on matters relating to the amplifier. And as I previously mentioned I have invested in a new sound card and software.
So a big thank you for your interest.
Good News
Despite apparent reservations about the use of inductors/transformers in audio amplifiers "I'm not in love with transformers, or any other part that has distortion, but sometimes they have their uses." (14th October 2004) see:
http://www.diyaudio.com/forums/showthread.php?postid=491741#post491741
The one and only Mr Nelson Pass has now implemented these in his latest Zen amplifier series "Zen Variations 7: More fun with Son of Zen and SuperSymmetry".
http://www.passdiy.com/pdf/zen-v7.pdf
Inductive Loading (from page 6) has the Zen-V7T and for which Mr Nelson Pass now plots distortion graphs up to 30 watts (at 1KHz about 3%).
At last someone is mentioning the effective additional power rail generation "Because the coupled inductors can swing much higher
voltage than the supply – in an idealized case, they would peak at twice the supply."
I am sure Mr Nelson Pass will add about the preference of using bi-filar or multi-filar windings to get close coupling between windings.
Mr Nelson Pass also shows a version, the Zen-V7E, for driving electrostatics on page 12.
(As an aside I note that electrostatic loudspeakers were in use before 1930!)
Mr Nelson Pass comments that he finds the sound of the Zen-V7T "lively and musical".
So all in all most excellent as I hope that this will encourage people to have a go at trying out transformer based amplifiers given that you can build mine and if you so desire change to the Zen topology with a few dabs of the soldering iron and a handful of different components.
And if you already have a Son of Zen version you can try out a bifilar wound toroid transformer very quickly.
Enjoy.
Best wishes,
Susan.
Hi everyone.
I see that this thread which I started on the 8th of October has passed the 28000 views and is now 6th out of 5890 threads for number of views in the solid state amplifier section (and seventh for number of replies).
It is a matter of some surprise to me the interest and level of discussion that has occurred and it is very pleasing to hear of people's progress with their versions of my design.
I do my best to answer questions as fully as possible but I am aware that in all the replies there must be some which I have inadvertently overlooked.
As it is the last few weeks I have probably been spending up to 50% of my available "work" time directly or indirectly on matters relating to the amplifier. And as I previously mentioned I have invested in a new sound card and software.
So a big thank you for your interest.
Good News
Despite apparent reservations about the use of inductors/transformers in audio amplifiers "I'm not in love with transformers, or any other part that has distortion, but sometimes they have their uses." (14th October 2004) see:
http://www.diyaudio.com/forums/showthread.php?postid=491741#post491741
The one and only Mr Nelson Pass has now implemented these in his latest Zen amplifier series "Zen Variations 7: More fun with Son of Zen and SuperSymmetry".
http://www.passdiy.com/pdf/zen-v7.pdf
Inductive Loading (from page 6) has the Zen-V7T and for which Mr Nelson Pass now plots distortion graphs up to 30 watts (at 1KHz about 3%).
At last someone is mentioning the effective additional power rail generation "Because the coupled inductors can swing much higher
voltage than the supply – in an idealized case, they would peak at twice the supply."
I am sure Mr Nelson Pass will add about the preference of using bi-filar or multi-filar windings to get close coupling between windings.
Mr Nelson Pass also shows a version, the Zen-V7E, for driving electrostatics on page 12.
(As an aside I note that electrostatic loudspeakers were in use before 1930!)
Mr Nelson Pass comments that he finds the sound of the Zen-V7T "lively and musical".
So all in all most excellent as I hope that this will encourage people to have a go at trying out transformer based amplifiers given that you can build mine and if you so desire change to the Zen topology with a few dabs of the soldering iron and a handful of different components.
And if you already have a Son of Zen version you can try out a bifilar wound toroid transformer very quickly.
Enjoy.
Best wishes,
Susan.
Hi peranders,
At the bottom of the page for the Solid State threads there is a search line with various options to present the listed topics in different orders.
Showing threads 1 to 25 of 5891, sorted by [views/replies/etc.] in [decending/asending] order, from [the beginning/last year/etc.]
Just select the version you want and click on the GO icon.
Hey presto!
I have not included the Announcements and the Permanent entries as they don't move
Best wishes,
Susan.
peranders said:Susan, you are obviously smarter than Ibecause I can't figure out how you have made these searches about top ten list of views? How did you do?
At the bottom of the page for the Solid State threads there is a search line with various options to present the listed topics in different orders.
Showing threads 1 to 25 of 5891, sorted by [views/replies/etc.] in [decending/asending] order, from [the beginning/last year/etc.]
Just select the version you want and click on the GO icon.
Hey presto!
I have not included the Announcements and the Permanent entries as they don't move
Best wishes,
Susan.
I like complicated solutions as some of you already know. "Less is more" is not my style more like "more is more". So far (sorry Nelson) I have had no desire to build (but I don't mind listening if the opportunity comes) any Pass amps but this extremely simple amp is something which have caught my interest, just becuase of it's simpleness. Maybe I'll built a Zeus amp. I have all the stuff already.
Re: WARNING!!!
and we are discussing this at Hififorum.nu. My suggestion is that the transformer needs some interfacing towards the mosfets. Maybe the capacitance's of the mosfets together with the properties of the transformer that are casusing troubles?
His amp is not optimalCirclomanen said:DO NOT OMIT THE GATE-SOURCE VOLTAGE PROTECTION!!!!!
I Played realy loud and all of a sudden it was quiet. Both fets was fully open and it had 28 ampere current running through the transformer....
No good.
Me:
allrightythen....I was playing Ricky Martin and my amp developed some kind of allergy.
and we are discussing this at Hififorum.nu. My suggestion is that the transformer needs some interfacing towards the mosfets. Maybe the capacitance's of the mosfets together with the properties of the transformer that are casusing troubles?
Re: similar design & transformers in b2 spice
Hi Mike,
Yes, sorry. My weekend has been quite busy and I am still playing catch-up. I was updating my website to make the audio stuff more accessible and domestic duties overtook me yesterday.
Thanks. I have tried to keep the audio path as simple and direct as possible.
Outside the path that carries the audio I have no issues with as much complexity as might be desired.
Temperature tracking mosfet bias circuits with a (PIC/Atmel) processor and ADC/DAC setup is fine as far as I am concerned (as long as it is properly screened).
My power protection circuit (still finishing this off) has almost as many parts as the actual amplifier.
For those that want to know what this possibly cryptic and leading comment might be about here is a quick overview of what I have in mind.
The basics for this the circuit uses yet another mosfet in the power rail which is connected up as a form of semi-regulator and it has an SCR on the gate to ground which is triggered by a sense circuit from a load resistor in the supply rail (after the storage capacitor(s)).
In normal operation the mosfet supplies a semi-regulated voltage at a level set by a zener less the gate/drain drop i.e. a 37 volt zener would regulate out to around 33 volts using IRFP140/150 type devices.
The advantage of this is that one can have a little higher supply voltage which allows for mains variations and some of the power what would otherwise get dissipated in the driver mosfets gets to go into the regulator part.
When the over-current condition is triggered the SCR which is across the zener fires and clamps the mosfet gate to ground. (One also had to take care of the normal gate/drain zener protection for this to ensure that the mosfet gate does not go out of bounds.)
Power is therefore removed from the driver mosfets and remains off until the amplifier is turned off and on again (after a pause to let the main caps bled out).
It would appear that many of the standard main toroids now have bifilar wound secondaries "to get better matching when used in parallel".
Otherwise you can wind your own bifilar choke easily enough. Buy the 200VA transformer kit and two reals of enameled wire from RS. Strip off the mains windings from the bobbin and re-wind. I don't know what quality of steal is used for the RS transformer laminations - I have asked but they didn't know.
I have been speaking with Sowter but a they don't have the smaller transformer size in stock I am looking at a version using the 120 size laminations. However I have to wind one up myself to ensure that it all works as I intend and I have not yet had the time to do this.
Transformers from Sowter will be well over a hundred pounds each, so if you can do this yourself or use a toroid then that would be a cheaper option to get you going.
To recap the "Rules"
1. Use a transformer of four times the mains VA rating for the number of watts you want to drive.
I.e. For 50 watts audio use a minimum of a 200VA transformer.
2. For an off the shelf toroid transformer using the secondaries only chose the secondary voltage to roughly match that of the power supply rail you are using.
E.g. If you have around 34 volts DC supply then use a 2 x 30 Vac toroid.
For those of us using RS (they do do credit card orders) the following would be suitable:
RS # 422-5348 ( 225 VA, 2 x 30 Vac, 3.75 Amps per winding.) at UKP 22
or for an extra three pounds
RS # 257-5231 ( 300 VA, 2 x 35 Vac, 4.28 Amps per winding.) at UKP 25.50
The latter might be preferable if you were using higher bias levels.
Both of these have dual 115/230 Vac primaries, which give a little more flexibility, however the single 230 Vac primary versions are four or five pounds cheaper.
I got stuck with the simulation as well. In practice the circuit isn't as hair trigger pass/fail as the simulation might suggest.
However good matching and coupling between windings is important, hence the multi-filar winding.
As you have probably noticed I am not keen on feedback amplifiers to drive speakers. In my opinion one can use feedback all the way through, but not for the very last stage that is driving the reactive and inconsistent load of the speakers.
My amplifier topology gives under 1% distortion even at 35 watts. It would only need very light feedback to drop this right down, but would be a step away from the simplicity of a few bits of wire and a mosfet or two.
However AB comparisons are always good, and simple circuits that encourage experimentation can only lead to greater understanding and hopefully enjoyment.
I would certainly be interested in what you do find with this
Thank you.
Best wishes,
Susan.
Hi Mike,
Yes, sorry. My weekend has been quite busy and I am still playing catch-up. I was updating my website to make the audio stuff more accessible and domestic duties overtook me yesterday.
mikelm said:
Thanks. I have tried to keep the audio path as simple and direct as possible.
Outside the path that carries the audio I have no issues with as much complexity as might be desired.
Temperature tracking mosfet bias circuits with a (PIC/Atmel) processor and ADC/DAC setup is fine as far as I am concerned (as long as it is properly screened).
My power protection circuit (still finishing this off) has almost as many parts as the actual amplifier.
For those that want to know what this possibly cryptic and leading comment might be about here is a quick overview of what I have in mind.
The basics for this the circuit uses yet another mosfet in the power rail which is connected up as a form of semi-regulator and it has an SCR on the gate to ground which is triggered by a sense circuit from a load resistor in the supply rail (after the storage capacitor(s)).
In normal operation the mosfet supplies a semi-regulated voltage at a level set by a zener less the gate/drain drop i.e. a 37 volt zener would regulate out to around 33 volts using IRFP140/150 type devices.
The advantage of this is that one can have a little higher supply voltage which allows for mains variations and some of the power what would otherwise get dissipated in the driver mosfets gets to go into the regulator part.
When the over-current condition is triggered the SCR which is across the zener fires and clamps the mosfet gate to ground. (One also had to take care of the normal gate/drain zener protection for this to ensure that the mosfet gate does not go out of bounds.)
Power is therefore removed from the driver mosfets and remains off until the amplifier is turned off and on again (after a pause to let the main caps bled out).
It would appear that many of the standard main toroids now have bifilar wound secondaries "to get better matching when used in parallel".
Otherwise you can wind your own bifilar choke easily enough. Buy the 200VA transformer kit and two reals of enameled wire from RS. Strip off the mains windings from the bobbin and re-wind. I don't know what quality of steal is used for the RS transformer laminations - I have asked but they didn't know.
I have been speaking with Sowter but a they don't have the smaller transformer size in stock I am looking at a version using the 120 size laminations. However I have to wind one up myself to ensure that it all works as I intend and I have not yet had the time to do this.
Transformers from Sowter will be well over a hundred pounds each, so if you can do this yourself or use a toroid then that would be a cheaper option to get you going.
To recap the "Rules"
1. Use a transformer of four times the mains VA rating for the number of watts you want to drive.
I.e. For 50 watts audio use a minimum of a 200VA transformer.
2. For an off the shelf toroid transformer using the secondaries only chose the secondary voltage to roughly match that of the power supply rail you are using.
E.g. If you have around 34 volts DC supply then use a 2 x 30 Vac toroid.
For those of us using RS (they do do credit card orders) the following would be suitable:
RS # 422-5348 ( 225 VA, 2 x 30 Vac, 3.75 Amps per winding.) at UKP 22
or for an extra three pounds
RS # 257-5231 ( 300 VA, 2 x 35 Vac, 4.28 Amps per winding.) at UKP 25.50
The latter might be preferable if you were using higher bias levels.
Both of these have dual 115/230 Vac primaries, which give a little more flexibility, however the single 230 Vac primary versions are four or five pounds cheaper.
I got stuck with the simulation as well. In practice the circuit isn't as hair trigger pass/fail as the simulation might suggest.
However good matching and coupling between windings is important, hence the multi-filar winding.
As you have probably noticed I am not keen on feedback amplifiers to drive speakers. In my opinion one can use feedback all the way through, but not for the very last stage that is driving the reactive and inconsistent load of the speakers.
My amplifier topology gives under 1% distortion even at 35 watts. It would only need very light feedback to drop this right down, but would be a step away from the simplicity of a few bits of wire and a mosfet or two.
However AB comparisons are always good, and simple circuits that encourage experimentation can only lead to greater understanding and hopefully enjoyment.
I would certainly be interested in what you do find with this
Great thread....
thanks
mike
Thank you.
Best wishes,
Susan.
Hi peranders ,
Which is why I encourage the use of cheap IRFPXXX parts to start with as well as a sensible quick blow fuse in the power rail AFTER the smoothing caps.
Blowing up a couple of IRFP140s is irritating but wiping out 50 pounds plus of matched pair lateral mosfets would bring me to tears
Also there is not uncertainly as to where the failure is, and it is quick to repair (we are all using fuses or a current limiting bench PSU aren't we).
And yes, I did blow up several sets of devices in the early days so I do have the T shirt with a picture of an IRF150 in a TO3 case upside down with bent pins and a wisp of magic smoke rising
Best wishes,
Susan.
peranders said:
Which is why I encourage the use of cheap IRFPXXX parts to start with as well as a sensible quick blow fuse in the power rail AFTER the smoothing caps.
Blowing up a couple of IRFP140s is irritating but wiping out 50 pounds plus of matched pair lateral mosfets would bring me to tears
Also there is not uncertainly as to where the failure is, and it is quick to repair (we are all using fuses or a current limiting bench PSU aren't we).
And yes, I did blow up several sets of devices in the early days so I do have the T shirt with a picture of an IRF150 in a TO3 case upside down with bent pins and a wisp of magic smoke rising
Best wishes,
Susan.
Power long-tailed pair.
There is another benefit too - you don't need to temperature compensate the gate bias voltage. If the current is set to X amps then that is where it stays.
Soooo...... here is what I suggest. In Susan's amplifier, it would be interesting to remove the temperature compensated bias circuit and instead put a constant current sink from the transformer centre tap to ground. The overall complexity would be similar anyway, plus you gain the extra benefit of forcing the mosfets to conduct in a complementary fashion no matter what.
P.S. It should improve power supply rejection even further too.
Someting I find interesting about that Zen amplifier is the use of a single constant current sink from the sources to earth. If the total current is very constant it means that as one mosfet increases say 100mA then the other side *has* to reduce by the same amount, so we gain a more-or-less exact complement of action between the two sides. This is not necessarily the case without the CCS with two imperfectly matched devices.Susan-Parker said:
There is another benefit too - you don't need to temperature compensate the gate bias voltage. If the current is set to X amps then that is where it stays.
Soooo...... here is what I suggest. In Susan's amplifier, it would be interesting to remove the temperature compensated bias circuit and instead put a constant current sink from the transformer centre tap to ground. The overall complexity would be similar anyway, plus you gain the extra benefit of forcing the mosfets to conduct in a complementary fashion no matter what.
P.S. It should improve power supply rejection even further too.
Re: Amp Changes 2 - running
Hi Uwe,
Thanks for the progress update
If I understand your configuration correctly you are now running with two separate and non connected output inductors in a dual bridged singled ended configuration.
Nice
With modern mechanized winding techniques and a high level of material control transformers of a same type and from the same batch are probably pretty close to identical.
You will need to run at a higher bias level then in the push-pull configuration. 91 dB sensitivity speakers helps a lot here.
Getting a better match between mosfets should help - I aim for 10 mV or better. However I have run with a lot more as well and it isn't as bad as one might expect.
Just make sure that the mosfets do have their gate protection as if one goes your speaker won't be happy.
Good that it is working well and that you like the sound.
Best wishes,
Susan.
Hi Uwe,
Thanks for the progress update
dx.master said:
If I understand your configuration correctly you are now running with two separate and non connected output inductors in a dual bridged singled ended configuration.
Nice
With modern mechanized winding techniques and a high level of material control transformers of a same type and from the same batch are probably pretty close to identical.
You will need to run at a higher bias level then in the push-pull configuration. 91 dB sensitivity speakers helps a lot here.
Getting a better match between mosfets should help - I aim for 10 mV or better. However I have run with a lot more as well and it isn't as bad as one might expect.
Just make sure that the mosfets do have their gate protection as if one goes your speaker won't be happy.
Good that it is working well and that you like the sound.
Best wishes,
Susan.
Re: Power long-tailed pair.
Hi Circlotron ,
Thank you for the suggestion.
I am not sure that it works as simply as that because I am using the mosfets as followers i.e. current amplifiers, and Mr Nelson Pass is using the mosfets as classic gain stages i.e. voltage amplifies.
If I was going to try something I would be looking at using a constant current source the other side - perhaps using a P channel complimentary mosfet to the driver devices.
This reduces the quiescent power dissipation in the drivers and could be incorporated as part of the power protection scheme.
However your sugestion is something that can be readily tried out with a few bits in short order - and it is always worth experimenting.
Thanks.
Best wishes.
Susan.
Hi Circlotron ,
Circlotron said:
Someting I find interesting about that Zen amplifier is the use of a single constant current sink from the sources to earth. If the total current is very constant it means that as one mosfet increases say 100mA then the other side *has* to reduce by the same amount, so we gain a more-or-less exact complement of action between the two sides. This is not necessarily the case without the CCS with two imperfectly matched devices.
There is another benefit too - you don't need to temperature compensate the gate bias voltage. If the current is set to X amps then that is where it stays.
Soooo...... here is what I suggest. In Susan's amplifier, it would be interesting to remove the temperature compensated bias circuit and instead put a constant current sink from the transformer centre tap to ground. The overall complexity would be similar anyway, plus you gain the extra benefit of forcing the mosfets to conduct in a complementary fashion no matter what.
P.S. It should improve power supply rejection even further too.
Thank you for the suggestion.
I am not sure that it works as simply as that because I am using the mosfets as followers i.e. current amplifiers, and Mr Nelson Pass is using the mosfets as classic gain stages i.e. voltage amplifies.
If I was going to try something I would be looking at using a constant current source the other side - perhaps using a P channel complimentary mosfet to the driver devices.
This reduces the quiescent power dissipation in the drivers and could be incorporated as part of the power protection scheme.
However your sugestion is something that can be readily tried out with a few bits in short order - and it is always worth experimenting.
Thanks.
Best wishes.
Susan.
Hi Circlotron,
Not to worry, and your thoughts are apreciated.
I know only too well how easy it is to miss things even on simple circuits and having input from people is always welcome.
Thanks.
Best wishes,
Susan.
Circlotron said:Drat! Your right. Basically you can't use a mosfet source as an output for signal and an input for a correction factor at the same time.
Not to worry, and your thoughts are apreciated.
I know only too well how easy it is to miss things even on simple circuits and having input from people is always welcome.
Thanks.
Best wishes,
Susan.
mosfet gate capacitance
Hi Peranders,
I have read through the posts - I didn't realize that the discussion had gone beyond the narrow margins of this DIYaudio thread
Remember that the gate capacitance of the mosfet that the input transformer sees is the gate to drain, not the gate to source. I.e. Crss not Ciss.
E.g. For a IRFP140 the Crss is below 400 pF overall and well below 200 pF for small signals.
This gets multiplied back though by the input transformer as far as the line driver is concerned.
So something that can cope with driving a capacitive load is needed - that TI headphone driver part seems to work okay here.
Best wishes,
Susan.
Hi Peranders,
peranders said:
I have read through the posts - I didn't realize that the discussion had gone beyond the narrow margins of this DIYaudio thread
Remember that the gate capacitance of the mosfet that the input transformer sees is the gate to drain, not the gate to source. I.e. Crss not Ciss.
E.g. For a IRFP140 the Crss is below 400 pF overall and well below 200 pF for small signals.
This gets multiplied back though by the input transformer as far as the line driver is concerned.
So something that can cope with driving a capacitive load is needed - that TI headphone driver part seems to work okay here.
Best wishes,
Susan.
Re: Re: Amp Changes 2 - running
Hi Susan-Parker,
That's right. The schematic on my website will/ show that.
All 3 IRF 530N (Mosfet) are match to 4,19 Volt.
see
What do you mean? The Gate Restistor and the one or two Z-Diode ?
I use to 10V Z-Diode.
see
Yes i, need more Outputpower and lower input voltage.
Now I drive the Zeus with my Behringer UB802 an the "Control Room Output" and a (old) CD Player from Marantz CD-36:
see Behringer UB802 .
thanke you for the comments.
- uwe
-= http://bauteile-fuer-die-elektronik.de/projekte/Zeus/ =-
Hi Susan-Parker,
Susan-Parker said:If I understand your configuration correctly you are now running with two separate and non connected output inductors in a dual bridged singled ended configuration.
Nice
With modern mechanized winding techniques and a high level of material control transformers of a same type and from the same batch are probably pretty close to identical.
That's right. The schematic on my website will/ show that.
All 3 IRF 530N (Mosfet) are match to 4,19 Volt.
see
An externally hosted image should be here but it was not working when we last tested it.
.
What do you mean? The Gate Restistor and the one or two Z-Diode ?
I use to 10V Z-Diode.
see
An externally hosted image should be here but it was not working when we last tested it.
.Yes i, need more Outputpower and lower input voltage.
Now I drive the Zeus with my Behringer UB802 an the "Control Room Output" and a (old) CD Player from Marantz CD-36:
see Behringer UB802 .
thanke you for the comments.
- uwe
-= http://bauteile-fuer-die-elektronik.de/projekte/Zeus/ =-
Re: Re: Re: Amp Changes 2 - running
Hi Uwe,
Excellent.
How many parts did it take to get three matched up?
Yes
Are you thinking of a 2 stage version of the ampifier (the input toroid TR1 you are using should have bifilar 12 volt windings)?
Or a line driver?
Very pleasing to see your progress.
Thanks for the updates.
Best wishes,
Susan.
Hi Uwe,
dx.master said:
Excellent.
How many parts did it take to get three matched up?
Yes
Are you thinking of a 2 stage version of the ampifier (the input toroid TR1 you are using should have bifilar 12 volt windings)?
Or a line driver?
Very pleasing to see your progress.
Thanks for the updates.
Best wishes,
Susan.
Balanced Line Drivers
Hi Uwe,
Okay, had a look at the Behringer UB802 specs.
You could add a balanced line driver using op amps, e.g. the OP275
http://hjem.get2net.dk/tkhifi/data-blade/op275.pdf
Figure 17, page 10.
Which is the part I used originally with my input transformer set for 150 or 600 ohm impedance. Whilst not everyone's cup of tea the op275 is specified to work into a 600 ohm load and seems to cope okay with 150 ohms.
For lower input impedances (and nice square wave test results) the current drive TI TPA6120 part works better but if one is using both sides to drive one transformer it is necessary to use the center tap of the 12V-0-12V returned to the driver's 0 volts as well otherwise the TPA6120 complains (big time).
See:
http://www.susan-parker.co.uk/zeus-input-line-driver-top-1.jpg
http://www.susan-parker.co.uk/zeus-input-line-driver-bottom-1.jpg
for pics of my TPA6120 tester built from samples and componets to hand. I now have some precision resistors to fit for the feedback bits.
It is being driven by the OP275 diferential driver which is optionally bandwidth limited to 200 kHz.
Best wishes,
Susan.
Hi Uwe,
Okay, had a look at the Behringer UB802 specs.
You could add a balanced line driver using op amps, e.g. the OP275
http://hjem.get2net.dk/tkhifi/data-blade/op275.pdf
Figure 17, page 10.
Which is the part I used originally with my input transformer set for 150 or 600 ohm impedance. Whilst not everyone's cup of tea the op275 is specified to work into a 600 ohm load and seems to cope okay with 150 ohms.
For lower input impedances (and nice square wave test results) the current drive TI TPA6120 part works better but if one is using both sides to drive one transformer it is necessary to use the center tap of the 12V-0-12V returned to the driver's 0 volts as well otherwise the TPA6120 complains (big time).
See:
http://www.susan-parker.co.uk/zeus-input-line-driver-top-1.jpg
http://www.susan-parker.co.uk/zeus-input-line-driver-bottom-1.jpg
for pics of my TPA6120 tester built from samples and componets to hand. I now have some precision resistors to fit for the feedback bits.
It is being driven by the OP275 diferential driver which is optionally bandwidth limited to 200 kHz.
Best wishes,
Susan.