Hello everyone,
I purchased two 70W output transformers from Antek since they were the cheapest I could find. I haven't read anything positive about them on this forum so I am a little worried how they will sound. I am an electronics engineering technology student and my capstone project is to design a hybrid tube amplifier using a solid state front end and a tube power amplifier using kt88 tubes. This is my first tube project.
After reading a thread about Antek's OT's, I realized there is not a lot of information on them. The specs given by Antek are pitiful to say the least. Since I have two of them, and access to good equipment (at school), I have decided to share my measurements. Attached is a pdf of my results. I still have to characterize the transformers a little bit more and do a proper frequency response test using a resistor (equal to the primary impedance) in series with the function generator. However, this information should give a good idea of how these transformers will behave.
The construction of these transformers seems very good. They are encased in heavy gauge steel and are quite heavy (~15lbs each). I characterized both of them to make sure they were electrically identical. If any of you want me to do further tests or have feedback please let me know. I would be glad to help out if anyone is interested in buying these transformers.
Todd
I purchased two 70W output transformers from Antek since they were the cheapest I could find. I haven't read anything positive about them on this forum so I am a little worried how they will sound. I am an electronics engineering technology student and my capstone project is to design a hybrid tube amplifier using a solid state front end and a tube power amplifier using kt88 tubes. This is my first tube project.
After reading a thread about Antek's OT's, I realized there is not a lot of information on them. The specs given by Antek are pitiful to say the least. Since I have two of them, and access to good equipment (at school), I have decided to share my measurements. Attached is a pdf of my results. I still have to characterize the transformers a little bit more and do a proper frequency response test using a resistor (equal to the primary impedance) in series with the function generator. However, this information should give a good idea of how these transformers will behave.
The construction of these transformers seems very good. They are encased in heavy gauge steel and are quite heavy (~15lbs each). I characterized both of them to make sure they were electrically identical. If any of you want me to do further tests or have feedback please let me know. I would be glad to help out if anyone is interested in buying these transformers.
Todd
Judging by the pinout they look to be push-pull transformers...
What did you use to measure the L and C? If you used an inductance/capacitance meter, I suggest that you ensure that it excites the DUT with a frequency within the audio range. 1 kHz would be my choice.
Also, note that you can get the leakage inductance by measuring the inductance of the primary with the secondary shorted.
15 nF for the primary. That sounds rather high...
~Tom
What did you use to measure the L and C? If you used an inductance/capacitance meter, I suggest that you ensure that it excites the DUT with a frequency within the audio range. 1 kHz would be my choice.
Also, note that you can get the leakage inductance by measuring the inductance of the primary with the secondary shorted.
15 nF for the primary. That sounds rather high...
~Tom
Hi Todd,
Well done and very informative! Maybe inductance is a little on the low side for a PP. Will not tell how it will sound though.
What´s there to worry about? There´s only one way to find out: Build and listen.
Well done and very informative! Maybe inductance is a little on the low side for a PP. Will not tell how it will sound though.
What´s there to worry about? There´s only one way to find out: Build and listen.
Each half of the primary turns ratio was measured to be 18.5 so the turns ratio from plate to plate is 37. With an 8 Ohm load, the reflected impedance is ~11kOhm plate to plate. The impedance seen by each tube is then 2783 Ohm.
I did a frequency response with a 2200 Ohm resistor in series on each half of the primary and the results are scary! See attached PDF.
I did use an LRC meter. I scoped the output of the meter and it measures
the capacitance at 1 kHz excitation, but the inductance is around 120 Hz excitation frequency. This is probably why the inductance values seem low.
I also took the two transformers and hooked the primaries together (Orange to Orange, Gray to Gray) and drove the 8 Ohm tap with an audio amplifier with a resistive load on the other 8 Ohm tap. The frequency response was a little bit better this way. Not sure what to expect though once these transformers are in a working amp.
Todd
I did a frequency response with a 2200 Ohm resistor in series on each half of the primary and the results are scary! See attached PDF.
I did use an LRC meter. I scoped the output of the meter and it measures
the capacitance at 1 kHz excitation, but the inductance is around 120 Hz excitation frequency. This is probably why the inductance values seem low.
I also took the two transformers and hooked the primaries together (Orange to Orange, Gray to Gray) and drove the 8 Ohm tap with an audio amplifier with a resistive load on the other 8 Ohm tap. The frequency response was a little bit better this way. Not sure what to expect though once these transformers are in a working amp.
Todd
A typical OT has more like 500 pF distributed primary capacitance. 15 nF would be 30 times as high. These apparently are just random wound xfmrs, like for 60 Hz power, which leads to the high capacitance. The second plot is what you will get in an amplifier with a 2200 Ohm source impedance.
The first plot is driven at very low impedance by the amplifier, and since the leakage L is very low for toroids, the second OT is also being driven with a low source Z. That overcomes the capacitance problem, but not the resonance problem appearing at 10 KHz. The resonance is at the frequency determined by L_leak and C_distributed. Low cost OTs usually have this resonance at 30 to 50 KHz. 10 KHz is way too low. A high quality OT will have the resonance up at 100 to 250 KHz.
The first plot is driven at very low impedance by the amplifier, and since the leakage L is very low for toroids, the second OT is also being driven with a low source Z. That overcomes the capacitance problem, but not the resonance problem appearing at 10 KHz. The resonance is at the frequency determined by L_leak and C_distributed. Low cost OTs usually have this resonance at 30 to 50 KHz. 10 KHz is way too low. A high quality OT will have the resonance up at 100 to 250 KHz.
Last edited:
Smoking-amp, what would you recommend for an output transformer. I have been looking at the Edcors since they are cheap and people seem to like them. I am using Kt88's in push pull and looking for a power output of 60W. The CXPP60-MS-6.6K OT seem like an OK choice for my needs. Your thoughts?
How much global feedback are you intending to use? Key issue!!
The CXPP-MS-6.6K, by my measurements, has a resonance around 30 KHz, placing it at the bottom end of the low cost HiFi OTs (but better than guitar OTs). Higher Ohm primaries would be even more marginal. This provides only enough bandwidth to support some minimal improvement in damping factor via global feedback.
Local feedback to reduce the source impedance is a must (Schade feedbacks, E linear, or using the secondary for some low % CFB). Also be aware that most of these low cost OTs are power rated using flux density all the way up to saturation (20Kgauss or more). A real Hi-Fi OT (E-I lam, but toroids can use closer to Sat. rating) rating would be at around 8000 or 9000 gauss. So if you want to still be HiFi at the 60 Watt level, then use the next size up OT. The HF end may likely be even more restricted though with a bigger OT.
I would suggest that you do the same measurements with any OT that you purchase, as you did with the Antek. DIYers perform so few tests on their OTs that almost anything passes for HiFi these days. Keep in mind that many practitioners of DIY tube amps cannot hear above 10 KHz anymore, and so cannot hear the missing bandwidth or HF distortion. (the overwhelming dependance on how it "sounds" reviews, without any testing, has led to steadily declining OT quality, "Golden" ears are more likely gold plated lead ears)
High global feedback amplifiers have virtually dissapeared from the scene because the OTs cannot support this now. I only know of one toroidal OT manufacturer that lists detailed info/specs on their xfmrs since they are truly first rate. Most of the others would prefer you didn't know. (I don't think they know themsives either in a lot of cases, they don't even have test equipment)
Trying to build a high global feedback amplifier nowadays will virtually guarantee bad results. Some special techniques like E-linear, output stage Schading, and Elliptron (low %CFB Circlotron) or Circlotron configuration can improve the HF performance of a marginal OT if applied correctly or allow the use of a low primary Z OT(Elliptron, Circlotron). (a high %CFB, by a conventional CFB winding, generally will require a high quality OT for good balance and coupling)
I don't know what your budget is for an OT, but if you should decide to look for a high performance OT, some cautions are in order. Some of the pricey OTs around with special magnetic alloys or Silver wires only offer minor improvements while not addressing the real problems. Just wallet draining PR mainly.
Low magnetizing current (at low freq.) is usually considered a plus, but if it's low enough will lead to a loss of the "tubey" sound for low freqs. (your choice here, some may call it increased detail, accuracy, others call it flat, lifeless, non tubey ) Permalloy core material (very expensive) can lower the magnetizing current dramatically, but at considerable loss of power capability. (really only makes sense for input xfmrs.) Using C-core or a toroid (usual M6 cheapo material) will lower the magnetizing current dramatically too (flux 100% in the preferred direction) and provide a significant boost in power capability besides.
High bandwidth can be achieved by using progressive wound toroids (not random wound, = awful), or using C-core with the windings on the outer legs (not combined on the inner leg), or using Long-E laminations with the winding spread out as a thin layer. (the key is long thin windings) These were the secret in some classic OTs. The long E lams are still available as constant V xfmr lams (M6), but no one seems to use them for audio. If you want to wind your own OT, this is the way to go, easy to wind, high performance, cheap. One manufacturer even sent me a 50 Lb box free sample of them.
Then there is the balance issue between the two sides of the P-P primary. The guitar OTs don't even consider this. Cheap "HiFi" OTs may use larger wire on the outer windings to lower the resistance for better matching, but do not address the difference in leakage L between inner and outer windings. This leads to the two sides of the OT having obvious differences in the HF response. High quality OTs will use a split bobbin with equal winding buildups (and primary crossovers between them) between the two bobbin sections. This will equalize the winding R and leakage L and provide better coupling between the two primary sides besides (approx. like the Mac bifialr scheme does).
The long-E lams can fit this scheme (dual/or split bobbin) beautifully when two standard bobbins will fit onto the long E tongue together. Thats what I use, easy to wind. The two bobbin C-core scheme (outside legs) can do the same. Toroids require special winding techniques and expensive progressive winding machines to do this, but certainly can.
The CXPP-MS-6.6K, by my measurements, has a resonance around 30 KHz, placing it at the bottom end of the low cost HiFi OTs (but better than guitar OTs). Higher Ohm primaries would be even more marginal. This provides only enough bandwidth to support some minimal improvement in damping factor via global feedback.
Local feedback to reduce the source impedance is a must (Schade feedbacks, E linear, or using the secondary for some low % CFB). Also be aware that most of these low cost OTs are power rated using flux density all the way up to saturation (20Kgauss or more). A real Hi-Fi OT (E-I lam, but toroids can use closer to Sat. rating) rating would be at around 8000 or 9000 gauss. So if you want to still be HiFi at the 60 Watt level, then use the next size up OT. The HF end may likely be even more restricted though with a bigger OT.
I would suggest that you do the same measurements with any OT that you purchase, as you did with the Antek. DIYers perform so few tests on their OTs that almost anything passes for HiFi these days. Keep in mind that many practitioners of DIY tube amps cannot hear above 10 KHz anymore, and so cannot hear the missing bandwidth or HF distortion. (the overwhelming dependance on how it "sounds" reviews, without any testing, has led to steadily declining OT quality, "Golden" ears are more likely gold plated lead ears)
High global feedback amplifiers have virtually dissapeared from the scene because the OTs cannot support this now. I only know of one toroidal OT manufacturer that lists detailed info/specs on their xfmrs since they are truly first rate. Most of the others would prefer you didn't know. (I don't think they know themsives either in a lot of cases, they don't even have test equipment)
Trying to build a high global feedback amplifier nowadays will virtually guarantee bad results. Some special techniques like E-linear, output stage Schading, and Elliptron (low %CFB Circlotron) or Circlotron configuration can improve the HF performance of a marginal OT if applied correctly or allow the use of a low primary Z OT(Elliptron, Circlotron). (a high %CFB, by a conventional CFB winding, generally will require a high quality OT for good balance and coupling)
I don't know what your budget is for an OT, but if you should decide to look for a high performance OT, some cautions are in order. Some of the pricey OTs around with special magnetic alloys or Silver wires only offer minor improvements while not addressing the real problems. Just wallet draining PR mainly.
Low magnetizing current (at low freq.) is usually considered a plus, but if it's low enough will lead to a loss of the "tubey" sound for low freqs. (your choice here, some may call it increased detail, accuracy, others call it flat, lifeless, non tubey ) Permalloy core material (very expensive) can lower the magnetizing current dramatically, but at considerable loss of power capability. (really only makes sense for input xfmrs.) Using C-core or a toroid (usual M6 cheapo material) will lower the magnetizing current dramatically too (flux 100% in the preferred direction) and provide a significant boost in power capability besides.
High bandwidth can be achieved by using progressive wound toroids (not random wound, = awful), or using C-core with the windings on the outer legs (not combined on the inner leg), or using Long-E laminations with the winding spread out as a thin layer. (the key is long thin windings) These were the secret in some classic OTs. The long E lams are still available as constant V xfmr lams (M6), but no one seems to use them for audio. If you want to wind your own OT, this is the way to go, easy to wind, high performance, cheap. One manufacturer even sent me a 50 Lb box free sample of them.
Then there is the balance issue between the two sides of the P-P primary. The guitar OTs don't even consider this. Cheap "HiFi" OTs may use larger wire on the outer windings to lower the resistance for better matching, but do not address the difference in leakage L between inner and outer windings. This leads to the two sides of the OT having obvious differences in the HF response. High quality OTs will use a split bobbin with equal winding buildups (and primary crossovers between them) between the two bobbin sections. This will equalize the winding R and leakage L and provide better coupling between the two primary sides besides (approx. like the Mac bifialr scheme does).
The long-E lams can fit this scheme (dual/or split bobbin) beautifully when two standard bobbins will fit onto the long E tongue together. Thats what I use, easy to wind. The two bobbin C-core scheme (outside legs) can do the same. Toroids require special winding techniques and expensive progressive winding machines to do this, but certainly can.
100$ each for an OT is within my budget, hence why I am looking at the Edcors. I wish I would have done some more research before I bought the AnTek OT's but I am new to this field. For the purpose of this project, I am not looking at producing the most hi-fidelity amplifier since I am a poor student, but I do want something that sounds good. The requirements for this tube amp is it must use a microprocessor to control the volume, tone (bass, mid, treble), speaker balance and input select. My group chose an audio processor IC (TDA7439) for the pre-amp. We are using a schematic I found in AudioXpress for a KT88 hybrid tube amp by Bill Christie. He uses the IC driver LM4702 to drive (and phase splitter) two KT88 tubes in PP. The amount of global feedback he uses is around 15dB.
I should have the prototype of the power amp done by next week, so I can test the AnTek OT and hear them. I built a Maida regulator set at 500V for the anodes with 50mA (~57 Vgk) of quiescent cathode current. From what I have read, these seems like a common setup for KT88 tubes.
Thank you for taking the time to explain OT specs. This will help me a lot for my future tube amp builds. Winding my own transformer sounds like a good idea. Are there kits available for this?
Thanks,
Todd
I should have the prototype of the power amp done by next week, so I can test the AnTek OT and hear them. I built a Maida regulator set at 500V for the anodes with 50mA (~57 Vgk) of quiescent cathode current. From what I have read, these seems like a common setup for KT88 tubes.
Thank you for taking the time to explain OT specs. This will help me a lot for my future tube amp builds. Winding my own transformer sounds like a good idea. Are there kits available for this?
Thanks,
Todd
The inductance should be measured at low frequency and at a number of different excitation levels to determine linearity. Measuring at 1khz often gives improper information since it is generally very close to the self resonance frequency between the inductance and winding capacitance which generally falls between 500hz and 2Khz.
Well below the resonant frequency (SRF) the inductance (OCL or open circuit inductance) dominates and meters give accurate results. As you approach the SRF the load becomes very high and purely resistive which gives most meters a false reading. Above the SRF, the load becomes capacitive in nature so attempting to measure the inductance ignores the dominant loading factor.
dave
"The amount of global feedback he uses is around 15dB."
For the conventional sense of global feedback, 15 dB is, I think, very likley to go unstable with the CXPP60-MS-6.6K with real speakers and cables. Maybe with just a resistor load. Of course, you can always roll off the amplifier's excess gain at some low frequency in the audio band to make it stable again, but then only the lower part of the audio band gets linearized as intended. Not really HiFi, but the bass will sound good.
I would think more like 6 dB of global fdbk, and put the rest in as local feedback from the primary or UL taps. (for some Op Amp configuration, a resistor network that scales the feedbacks from multiple sources) That way the tube gain gets linearized, and the output impedance gets lowered for decent damping, without stability problems due to the OT. (also the OT magnetizing current gets handled better by the low Zout too.) To do "real" global feedback, the OT has to have excess bandwidth to support the correction loop with speed and stability. The cheap OTs just don't have it, they just pass the audio band.
The Antec toroid OT might work as a bass channel if it has more inductance than you measured. Try Dave's suggestion to measure it at 60 Hz or 120 Hz. The inductance goes up dramatically for M6 steel if you use more voltage than a typical LC meter uses. Try 6.3 VAC on the secondary, but watch out for the High Voltage on the primary then. (measure the secondary current)
I'm not aware of any OT winding kits, but there is a place that supplies the bobbins, tape etc.
Transformer Components
You can also just Google the net for bobbin and lamination sellers, but most of these deal only in huge quantity orders. You might be able to get a sample though.
Thomas and Skinner - Transformer Laminations
Cosmo Corporation - Coil Bobbin Catalog Main Menu
Foremost Plastic | Standard Bobbins | Plastic Injection
note:
With an Op Amp driver (single feedback insertion point) there could still be a stability problem, even with the scaled local feedbacks, since only one side of the OT is driven at a time, typically, for class AB. This means that 1/2 the feedback is still having to go thru the xfmr. Usually these local feedbacks are arranged separately for each output tube so that only the active tube and it's active feedback source have control at any given time. The Elliptron and Circlotron output configurations get around this problem by driving both local feedback takeoffs by each tube, but power supply and grid drive amplitude complications arise there.
For the conventional sense of global feedback, 15 dB is, I think, very likley to go unstable with the CXPP60-MS-6.6K with real speakers and cables. Maybe with just a resistor load. Of course, you can always roll off the amplifier's excess gain at some low frequency in the audio band to make it stable again, but then only the lower part of the audio band gets linearized as intended. Not really HiFi, but the bass will sound good.
I would think more like 6 dB of global fdbk, and put the rest in as local feedback from the primary or UL taps. (for some Op Amp configuration, a resistor network that scales the feedbacks from multiple sources) That way the tube gain gets linearized, and the output impedance gets lowered for decent damping, without stability problems due to the OT. (also the OT magnetizing current gets handled better by the low Zout too.) To do "real" global feedback, the OT has to have excess bandwidth to support the correction loop with speed and stability. The cheap OTs just don't have it, they just pass the audio band.
The Antec toroid OT might work as a bass channel if it has more inductance than you measured. Try Dave's suggestion to measure it at 60 Hz or 120 Hz. The inductance goes up dramatically for M6 steel if you use more voltage than a typical LC meter uses. Try 6.3 VAC on the secondary, but watch out for the High Voltage on the primary then. (measure the secondary current)
I'm not aware of any OT winding kits, but there is a place that supplies the bobbins, tape etc.
Transformer Components
You can also just Google the net for bobbin and lamination sellers, but most of these deal only in huge quantity orders. You might be able to get a sample though.
Thomas and Skinner - Transformer Laminations
Cosmo Corporation - Coil Bobbin Catalog Main Menu
Foremost Plastic | Standard Bobbins | Plastic Injection
note:
With an Op Amp driver (single feedback insertion point) there could still be a stability problem, even with the scaled local feedbacks, since only one side of the OT is driven at a time, typically, for class AB. This means that 1/2 the feedback is still having to go thru the xfmr. Usually these local feedbacks are arranged separately for each output tube so that only the active tube and it's active feedback source have control at any given time. The Elliptron and Circlotron output configurations get around this problem by driving both local feedback takeoffs by each tube, but power supply and grid drive amplitude complications arise there.
Last edited:
Using a HV transformer, variac and an ammeter on the primary is an easy and accurate (enough) way to plot inductance vs. excitation voltage.
When plotted you should see increasing inductance until saturation begins at which point it will drop again.
When plotted you should see increasing inductance until saturation begins at which point it will drop again.
- Status
- Not open for further replies.