The inductance fluctuate with frequency AND signal strenght.
I think you better measure frequency response for low, medium and large signals to know how the transformer performes. All preferable with the amplifier.
A 10kHz squarewave would be nice too.
I think you better measure frequency response for low, medium and large signals to know how the transformer performes. All preferable with the amplifier.
A 10kHz squarewave would be nice too.
Last edited:
There are a whole range of tests you could do on a transformer but most of them would be pointless without first knowing basic parameters like dc resistance and the inducatance at 50 or 60Hz neither of which is commonly published by manufacturers. The purpose of the document was to collect this basic information about a range of commonly used audio transformers to help people make an initial selection. Individuals could then supplement that with extended tests of their own.
Cheers
ian
Cheers
ian
The dc resistance has a constant value, inductance of the most transformers not.
I recently measured a line transformer who has a "healthy" -3dB point of 8Hz, probebly enough induction if i did a standard test. However, at lower signals the -3dB point shifted to 40Hz.......
Standard induction tests is not enough.
I recently measured a line transformer who has a "healthy" -3dB point of 8Hz, probebly enough induction if i did a standard test. However, at lower signals the -3dB point shifted to 40Hz.......
Standard induction tests is not enough.
There are a whole range of tests you could do on a transformer but most of them would be pointless without first knowing basic parameters like dc resistance and the inducatance at 50 or 60Hz neither of which is commonly published by manufacturers. The purpose of the document was to collect this basic information about a range of commonly used audio transformers to help people make an initial selection. Individuals could then supplement that with extended tests of their own.
Cheers
ian
Cheers
ian
In practice that statement is not true. There is an excellent correlation between achieved performance and these simple measurements.
Cheers
Ian
Cheers
Ian
Knowing everything is always nice but never possible. Knowing something is generally what we have to design to. Knowing nothing means design by braille.
I'd rather have parameters measured under published conditions than no parameters at all. Expecting the manufacturer to provide each individual customer with a complete set of measurements under the conditions specified by the customer is not realistic.
~Tom
I'd rather have parameters measured under published conditions than no parameters at all. Expecting the manufacturer to provide each individual customer with a complete set of measurements under the conditions specified by the customer is not realistic.
~Tom
Edcor made me 10K primary transformers for my Pyramid amps that run from 800V B+, 100W rated output, no problems at all.
Look at the both Tango transformers in this test:
SACThailand
The U808 has far more inductance then the XE-20S but in the frequency test you don't see any differance.
Un title page
SACThailand
The U808 has far more inductance then the XE-20S but in the frequency test you don't see any differance.
Un title page
Those tests have been around for years... pretty much to "technically showcase" their iron against the competition. The first link you're likely trying to show is this one, which is a frame page:
SAC Thailand
You might also want to download the Tango catalog and take a closer look at the specifications for each of the OPTs you're attempting a comparison of:
http://www.tubebooks.org/file_downloads/tango_tamura/Tango_Transformer.pdf
You'll notice that the specified inductance of the XE-20S and U-808 are specified at different DC current levels and impedance taps, so per the spec sheet, the difference in inductance isn't nearly as great as you imply as the inductance change vs DC current change is not going to be linear. Also, if you download the separate spec sheets for the XE20S and U-808, the manufacturer frequency response graphs are quite different between the two and they are different core construction as well.
Getting back to the referenced test, ALL transformers were measured at the same DC current level and same output drive level, so I wouldn't expect to see any real difference as neither OPT is near it's power limit. I also suspect some of those graphs were "edited a bit" so I don't have the same faith in them as you appear to.
Regards, KM
SAC Thailand
You might also want to download the Tango catalog and take a closer look at the specifications for each of the OPTs you're attempting a comparison of:
http://www.tubebooks.org/file_downloads/tango_tamura/Tango_Transformer.pdf
You'll notice that the specified inductance of the XE-20S and U-808 are specified at different DC current levels and impedance taps, so per the spec sheet, the difference in inductance isn't nearly as great as you imply as the inductance change vs DC current change is not going to be linear. Also, if you download the separate spec sheets for the XE20S and U-808, the manufacturer frequency response graphs are quite different between the two and they are different core construction as well.
Getting back to the referenced test, ALL transformers were measured at the same DC current level and same output drive level, so I wouldn't expect to see any real difference as neither OPT is near it's power limit. I also suspect some of those graphs were "edited a bit" so I don't have the same faith in them as you appear to.
Regards, KM
Tango didn't use a real tube for there test but a resistor/choke combination.
Not a real world measurement, even far from that as you compare the results with a real tube.
Sacthailand could do a better test but it is much better then the Tango test or just an inductance meter. If they tested also on a lower output level it would be more interesting.
Why you say: "ALL transformers were measured at the same DC current level and same output drive level, so I wouldn't expect to see any real difference as neither OPT is near it's power limit" ?
Why you say they edited the test results?
Not a real world measurement, even far from that as you compare the results with a real tube.
Sacthailand could do a better test but it is much better then the Tango test or just an inductance meter. If they tested also on a lower output level it would be more interesting.
Why you say: "ALL transformers were measured at the same DC current level and same output drive level, so I wouldn't expect to see any real difference as neither OPT is near it's power limit" ?
Why you say they edited the test results?
Last edited:
Exactly how do you know how Tango (or any of the other manufacturers) did their testing and determined it's not a realistic test?
Take a closer look at the tests you referenced:
1- No tube type listed, no bypass capacitor frequency calculation shown. This affects the low frequency roll-off considerably. For measurements, you would be better off using fixed-bias vs self-bias as it eliminates this from skewing the measurements. The same applies to the output tube... no one tube type will be optimal with varying loads from 1.9K to 5K.
2- Incomplete measurements on the SE-40 OPT (and the JS-6123S)... did they get bored, lazy or have an equipment failure? As it's an incomplete set of measurements, I would remove from the tests as such.
3- Look at the SE-40 graph, pretty sharp angle at the LF corner, I have not seen an OPT in real operation do that.
4- They specifically mention the XE-20S and U-808 in the description of their S-325 OPT but just stick to basic comments on the rest, sans the James unit. In most of the world, that's called marketing, as Tango have (had) the better reputation compared to the other brands shown.
I have done some measurements using a 45 DHT triode amplifier using the following OPTs. Note that only the OPT was swapped out, no other changes.
- Electra-Print (custom for the 45) (5K load)
- Hammond 125ESE (5K load)
- Hashimoto H-507S (5K load)
- James HS-6113S (5K load)
- Transcendar TT-012-OT (5K load)
All frequency response measurements were done at 1-watt output. With little exception, all did very similar on the LF response, HF varied quite a bit, which is expected and square response varied a lot between them.
I'm not saying their OPTs aren't good.... just that the tests had a goal versus a level-playing field. You can draw your own conclusions, I'll stick with mine.
Regards, KM
Take a closer look at the tests you referenced:
1- No tube type listed, no bypass capacitor frequency calculation shown. This affects the low frequency roll-off considerably. For measurements, you would be better off using fixed-bias vs self-bias as it eliminates this from skewing the measurements. The same applies to the output tube... no one tube type will be optimal with varying loads from 1.9K to 5K.
2- Incomplete measurements on the SE-40 OPT (and the JS-6123S)... did they get bored, lazy or have an equipment failure? As it's an incomplete set of measurements, I would remove from the tests as such.
3- Look at the SE-40 graph, pretty sharp angle at the LF corner, I have not seen an OPT in real operation do that.
4- They specifically mention the XE-20S and U-808 in the description of their S-325 OPT but just stick to basic comments on the rest, sans the James unit. In most of the world, that's called marketing, as Tango have (had) the better reputation compared to the other brands shown.
I have done some measurements using a 45 DHT triode amplifier using the following OPTs. Note that only the OPT was swapped out, no other changes.
- Electra-Print (custom for the 45) (5K load)
- Hammond 125ESE (5K load)
- Hashimoto H-507S (5K load)
- James HS-6113S (5K load)
- Transcendar TT-012-OT (5K load)
All frequency response measurements were done at 1-watt output. With little exception, all did very similar on the LF response, HF varied quite a bit, which is expected and square response varied a lot between them.
I'm not saying their OPTs aren't good.... just that the tests had a goal versus a level-playing field. You can draw your own conclusions, I'll stick with mine.
Regards, KM
The Tango measurement set-up is often published in there own datasheets (frequency response)
Yes, It would be better if SacThailand publiced all components in the test set-up but to compare it is good enough for the high frequencies but the low end is a bit tricky. It's the worse for the Hammond. Saying this, the inductance test is anyway rubbish because they measure only with ac signals and not preloaded with a dc current. These transformers are for Single Ended!
It's not importand that some totally unimportand data is missing. We want to see the frequency respons and to see where the resonance frequenties are and specially how well damped. The James and Audionote have not so good hf response and the Hammond is rubbish.
Some more importand data is missing too, Rdc sec, or the copper losses.
I miss the resonse 100kHz-1MHz. Maybe there are problems at higher frequenties too? We don't know now ( but some square waves indicate that some transformers have ....)
If they took some more samples the graphs would be a lot smoother (lf of the SE-40)
It's not the best test, much can be improved but it's a start.
Yes, It would be better if SacThailand publiced all components in the test set-up but to compare it is good enough for the high frequencies but the low end is a bit tricky. It's the worse for the Hammond. Saying this, the inductance test is anyway rubbish because they measure only with ac signals and not preloaded with a dc current. These transformers are for Single Ended!
It's not importand that some totally unimportand data is missing. We want to see the frequency respons and to see where the resonance frequenties are and specially how well damped. The James and Audionote have not so good hf response and the Hammond is rubbish.
Some more importand data is missing too, Rdc sec, or the copper losses.
I miss the resonse 100kHz-1MHz. Maybe there are problems at higher frequenties too? We don't know now ( but some square waves indicate that some transformers have ....)
If they took some more samples the graphs would be a lot smoother (lf of the SE-40)
It's not the best test, much can be improved but it's a start.
Last edited:
This is why I'm somewhat suspect of their tests and conclusions (which obviously show their OPT in a better light than the others), the tests are inconsistent/incomplete from OPT to OPT. Also, their Quality Factor is a joke... a ratio of what they measured for what purpose?? The Hammond 1628SE (as rubbish as it is) has a higher Quality Factor than the Tango XE-20S, so that couldn't possibly be a more meaningless representation, albeit their OPT scored the highest.
When you consider the wide variation in primary resistance from a low of 114 ohms to a high of 439 ohms, that alone contributes to a DC loss from just under 8 volts to over 30 volts from the plate supply at the reference current of 70ma. Was this compensated for? Yet another unknown.
The actual conditions for measuring the inductance is unknown, so I don't think it can be assumed that they used a standing DC current or not. Based on the Tango data (and the Hashimoto published data) the inductance is measured with with a standing DC current, which is correct for a single-ended OPT.
I guess you can call the test a start, but as it's not consistent between the various units, which in themselves are not the same impedance (hence an apples to oranges comparison), I have to discount the tests as more marketing than factual.
On the other hand, we've sorta strayed away from the OP's initial topic... oops.
Regards, KM
When you consider the wide variation in primary resistance from a low of 114 ohms to a high of 439 ohms, that alone contributes to a DC loss from just under 8 volts to over 30 volts from the plate supply at the reference current of 70ma. Was this compensated for? Yet another unknown.
The actual conditions for measuring the inductance is unknown, so I don't think it can be assumed that they used a standing DC current or not. Based on the Tango data (and the Hashimoto published data) the inductance is measured with with a standing DC current, which is correct for a single-ended OPT.
I guess you can call the test a start, but as it's not consistent between the various units, which in themselves are not the same impedance (hence an apples to oranges comparison), I have to discount the tests as more marketing than factual.
On the other hand, we've sorta strayed away from the OP's initial topic... oops.
Regards, KM
The initial topic was data from a specific manufacture.
I was trying to say that data as inductance, leakage, (winding) capacity measured with LCR meters don't tell anything about the actual transformer
Also SacThailand makes this mistake but luckely they did something more important: real data measured with a tube, a frequency response (but only at 1 output level) and a square wave.
Most other manufactures don't give data at all or give meaningless data.
A correction for dc losses is not needed. The differance is very very low.
I was trying to say that data as inductance, leakage, (winding) capacity measured with LCR meters don't tell anything about the actual transformer
Also SacThailand makes this mistake but luckely they did something more important: real data measured with a tube, a frequency response (but only at 1 output level) and a square wave.
Most other manufactures don't give data at all or give meaningless data.
A correction for dc losses is not needed. The differance is very very low.
Again more an addition to transformer specs than the OP's original question, but since measuring methods are part of that question:
I do not see equivalent capacitance listed as a necessary spec. Yet often that capacitance has a greater influence on h.f. response than leakage inductance. I am also not sure of the value in this case of a square wave response. It will show some resonance somewhere - we have L and C after all - but not where.
If anything I rather prefer a frequency response test under working conditions up to at least beyond any h.f. resonances. Then, as inductance is lowest at low exitation, the standard of measuring (at least for British and European countries) primary inductance at 5Vrms and 50 Hz seems to have been adopted since the days of the first high fidelity designs by Williamson and Leak. For the Williamson this represents a level of 2,5mW, a rather low one. Leakage reactance is usually shown at 1 kHz. This might have more comparison than practical value, but one learns to apply it and the measuring conditions are at least standard.
I will refrain from comment regarding limited available specs. It is almost like selling resistors without wattage specs or capacitors without voltage ratings. There appears to be satisfaction with Edcor procucts, if not with the electronic wherewithall of the agents. As far as I am concerned inductance and leakage inductance as above, equivalent capacitance, d.c. primary resistance (for OPTs), maximum power and a frequency graph under working conditions are a minimum - heavens, it is not as if these parameters take rocket science ability to measure. Voltage insulation: It can be measured fairly easily with a non-destructive ionometer. In addition, some sort of distortion figure would also help. I often find all these with reputable brands (I have seen such specs from both Sowther and Lundahl; I am not familiar with products from the East).
I do not see equivalent capacitance listed as a necessary spec. Yet often that capacitance has a greater influence on h.f. response than leakage inductance. I am also not sure of the value in this case of a square wave response. It will show some resonance somewhere - we have L and C after all - but not where.
If anything I rather prefer a frequency response test under working conditions up to at least beyond any h.f. resonances. Then, as inductance is lowest at low exitation, the standard of measuring (at least for British and European countries) primary inductance at 5Vrms and 50 Hz seems to have been adopted since the days of the first high fidelity designs by Williamson and Leak. For the Williamson this represents a level of 2,5mW, a rather low one. Leakage reactance is usually shown at 1 kHz. This might have more comparison than practical value, but one learns to apply it and the measuring conditions are at least standard.
I will refrain from comment regarding limited available specs. It is almost like selling resistors without wattage specs or capacitors without voltage ratings. There appears to be satisfaction with Edcor procucts, if not with the electronic wherewithall of the agents. As far as I am concerned inductance and leakage inductance as above, equivalent capacitance, d.c. primary resistance (for OPTs), maximum power and a frequency graph under working conditions are a minimum - heavens, it is not as if these parameters take rocket science ability to measure. Voltage insulation: It can be measured fairly easily with a non-destructive ionometer. In addition, some sort of distortion figure would also help. I often find all these with reputable brands (I have seen such specs from both Sowther and Lundahl; I am not familiar with products from the East).
I Use Them
Hi Everyone, an interesting discussion. I don't want to start a row here and I certainly respect the opinions of others whether the same or differing from mine. I use Edcor power, output and inter stage matchers all the time. None have failed to perform excellently. I specify them in all the commercial products a company of mine produces as well (we use 100's). I find the folks at Edcor to be really great and especially helpful when I want something designed and wound that is non standard. So....it brings me around to what I feel is a point we have skirted. As DIYERS are you more interested in specs and Spice models or more interested in sound? For me sound beats out all the specs and number crunching. Sure I use some in early design stages. Even when armed with all the data the results are often different to some degree than expected. I find that most of the time the end product needs a pretty good amount of bench testing, tweaking and listening to achieve the desired level of performance. It may be that we are still not able to model tube gear as well as we would like. One thing though that has not ever caused me problems are the transformers Edcor supplies. They have always managed to exceed the required parameters.
Hi Everyone, an interesting discussion. I don't want to start a row here and I certainly respect the opinions of others whether the same or differing from mine. I use Edcor power, output and inter stage matchers all the time. None have failed to perform excellently. I specify them in all the commercial products a company of mine produces as well (we use 100's). I find the folks at Edcor to be really great and especially helpful when I want something designed and wound that is non standard. So....it brings me around to what I feel is a point we have skirted. As DIYERS are you more interested in specs and Spice models or more interested in sound? For me sound beats out all the specs and number crunching. Sure I use some in early design stages. Even when armed with all the data the results are often different to some degree than expected. I find that most of the time the end product needs a pretty good amount of bench testing, tweaking and listening to achieve the desired level of performance. It may be that we are still not able to model tube gear as well as we would like. One thing though that has not ever caused me problems are the transformers Edcor supplies. They have always managed to exceed the required parameters.
Hi Bruce,
Neither do I want to re-start something which has been thoroughly discussed even to the extent of boredom - but you will agree that one must - er - start somewhere? One cannot just take a bag full of components and start thumb-sucking. Certain things need to be calculated if one wants even a slim chance of avoiding blowing up something expensive at an early stage!
Listening is an end operation, but what about the way there? Would it surprise you to know that many of the big designers never listened to their products until after they were finished? (Such is the consistent predictability of audio design and has it been for decades.) I am odd-man-out here as I am a professional EE and thus need some predictability in my designs if I want to avoid smoke coming out at an early stage or wasting money.
Regarding transformers then, from the evidence no one doubts the success of Edcor - it is the lack of information from the manufacturers which amazes. How does one avoid the situation that eg. when the bass is insufficient - keep on buying expensive OPTs until an acceptable result is reached - or get the right one from the beginning judged by its specs?
To that end then does one need specs of components ....
And apology for OT; just replying to the previous poster's remarks. Back to topic.
Neither do I want to re-start something which has been thoroughly discussed even to the extent of boredom - but you will agree that one must - er - start somewhere? One cannot just take a bag full of components and start thumb-sucking. Certain things need to be calculated if one wants even a slim chance of avoiding blowing up something expensive at an early stage!
Listening is an end operation, but what about the way there? Would it surprise you to know that many of the big designers never listened to their products until after they were finished? (Such is the consistent predictability of audio design and has it been for decades.) I am odd-man-out here as I am a professional EE and thus need some predictability in my designs if I want to avoid smoke coming out at an early stage or wasting money.
Regarding transformers then, from the evidence no one doubts the success of Edcor - it is the lack of information from the manufacturers which amazes. How does one avoid the situation that eg. when the bass is insufficient - keep on buying expensive OPTs until an acceptable result is reached - or get the right one from the beginning judged by its specs?
To that end then does one need specs of components ....
And apology for OT; just replying to the previous poster's remarks. Back to topic.
I wonder where they get their "virgin copper magnet wire" ?
Isn't extra virgin even better ? 🙂
boB
Isn't extra virgin even better ? 🙂
boB
Yes it is but you have to pay extra 🙂
Actually there wire is very nice documented:
https://www.edcorusa.com/t/mtr-magwire
Actually there wire is very nice documented:
https://www.edcorusa.com/t/mtr-magwire
Last edited:
- Status
- Not open for further replies.