Audio Transformer Inductance Ian Thompson-Bell January 2012
Introduction There seems to be a lot of mystery surrounding audio transformers which is not reduced by the reluctance of manufacturers to publish detailed specifications. This makes it hard for end users to choose between supposedly identical transformers from different manufacturers. When there is a significant price difference between them there seems to be little obvious justification for choosing the higher priced one other than reputation or hearsay concerning the 'sound' of a particular transformer. Given that most transformer manufacturers are equally well versed in the design and manufacture of transformers, the only real factors that can significantly affect the price of a transformer are the quality and quantity of the materials used in its construction. The two main materials used in transformers are wire for the windings and various materials for the magnetic core. If you use a bigger core and wind more turns of wire on it your transformer will be more expensive. This is interesting because using a bigger core and more turns of wire results in a larger inductance and the inductance of a transformer is a very important audio parameter that manufacturers seldom specify. The reason a transformer's inductance is important is that the primary winding inductance largely determines the low frequency response of the transformer. Suppose we have a 10K/600 transformer with a 600 ohm load connected across the secondary. This is reflected by the transformer to the primary where it 'looks like' 10K ohms so the driving amplifier only has to drive a 10K load which was basically the whole point of putting the transformer in. However, the reflected load is not the whole story because the inductive reactance of the primary winding is in parallel with it and the inductive reactance depends on the inductance. Suppose the transformer has a primary inductance of 20 Henries. The inductive reactance is proportional to frequency and is given by 2 x pi x f x L where L is the inductance in Henries and f is the frequency in Hertz. So at 1KHz the inductive reactance of our 20 Henry primary is over 120K which is clearly not a problem. However, at 20Hz, this reduces to around 2500 ohms which is much less than the advertised 10K ohms. If the driving circuit has a low enough output impedance it will be able to cope with this load without producing distortion at low frequencies and and with negligible loss of level. However, if the drive circuit is a tube one with an output impedance of 1500 ohms or more (a typical value) two things are likely to happen. First, the distortion at low frequencies is likely to rise considerably and secondly the response at 20Hz will be about 2dB down. Now, if the primary inductance was raised to 100 Henries (which needs more wire and probably a larger core) the inductive reactance at 20Hz would be well over 12K ohms and the tube driver would be perfectly happy. So, the bottom line is that a larger primary inductance gives a better low frequency
response and places little or no additional load on the driving circuit at low frequencies. With this in mind I recently invested in an Agilent U1371C LCR meter and used it to measure the primary and secondary inductances of a number of audio transformers. The results are given in following pages. For each one the measured primary and secondary inductance are given along with the calculated inductive reactance at 20Hz (XL20). The meter can read inductance at 100Hz, 120Hz and 1KHz. Without exception, the highest inductance reading was obtained at the lowest frequency (100Hz) and it is the 100Hz measurement that is quoted in the results section. After posting the initial results at groupdiy.com, several members offered their own measurements of other audio transformers. These results have now been appended.
Results Sowter 7490e microphone input transformer: This is Sowters' version of the 1:10 ratio mic input transformer for the Helios Type 69 mic pre. Primary L Secondary L Primary XL20 XL20 6H 619H 754 78K Note that the inductances are in the ratio of the turns ratio squared. The primary inductance is sufficient to give little primary attenuation at 20Hz from a 150 ohm source (0.16dB). The secondary inductance is enormous. Edcor XSE 10-50-8K output transformer: This transformer is intended for single ended tube amplifiers to drive a 50 ohm load reflected to 8K in the primary. I used this in a tube headphones amplifier design as a low cost alternative to the Sowter 8665. 5.5H 36mH 691 4.5 This is supposed to present an 8K ohm primary impedance but with only 5.5 Henries primary inductance its impedance only reaches 8k ohms at 230Hz. One reason the primary inductance is so low is that this transformer is gapped because it expected to carry dc anode current.. Note that Sowter quote the primary inductance of the 8665 at over 130 Henries which has an XL20 values of over 15K ohms. Carnhill VTB8291 9K6:600 output transformer: This is a standard output transformer intended for capacitor coupled tube output circuits. It is ungapped so it cannot carry dc current. 106H 11H 13K3 1382 The primary inductance is so high that there is negligible loading at 20Hz even for a driving source impedance as high as 2K ohms.
Carnhill VTB2281 600:600 output transformer: This is essentially the 600:600 version of the VTB8291. 12.7H 12.7H 1596 1596 Similar inductance to the 600 winding of the VTB8291.BBC LL/88 ASC 600:600 This is much smaller than its Carnhill equivalent and is contained in a cylindrical housing similar to those used by Sowter. I am not certain if this is intended as an input or output transformer. 20.6H 20.6H 2592 2592 This is a very well specified transformer with a more than adequate primary inductance. Even if fed from a 600 ohm source, the primary inductance is high enough that the loss at 20Hz is only 0.23dB. BBC LL81 10K:10K I believe this is a 10K bridging transformer so it expects a driving source impedance not greater than 600 ohms. Its construction is similar to the LL/88 but in a smaller cylindrical housing. 91H 91H 11K4 11K4 Another well specified transformer that presents a negligible load to a 600 ohm source. ROH DT-171 600:600 output This is a small PCB mounting open frame transformer. 14H 14H 1728 1728 Another well specified transformer with a more than adequate primary inductance.
Data Supplied by AZ999 Cinemag CMMI-7(C) Data supplied by Cinemag to AZ999. The is a nominal 200 ohm input microphone transformer with a 1:7 ratio and a 50 ohm primary tap. Recommended secondary load of 97.6K ohms represents a reflected primary load 2K ohms Primary L Primary XL20 30H 3768 The high primary inductance means the its inductive reactance is always large resulting in both minimal attenuation and phase shift at the lowest frequencies. Data Supplied by MathisD OEP Z3003e primaries in series for 1:1. This is a 10K:10K bridging transformer intended for either input or output applications. Using handheld meter: primaries 65.5H, secondaries 65.5H Using HP 4262A meter: at 120Hz - 82.6H at 1k - 45H at 10k - no reading, it will only read up to 10H at 10k. OEP Mic input 1:6 model Z21808 Using handheld meter: Primaries 5.83H Using HP 4262A meter: Primaries at 120Hz - 10.5H, at 1K - 6.9H Secondaries at 120Hz - 318H Some others measured with handheld meter Beyer 1:10 Mic input :Primaries - 5.43H
Beyer 1:15 Mic input (tiny sized canned model): Primaries 4.38H Lundahl LL1540 Primaries - no reading Secondaries - 117.6H and 117.3H for each winding Jensen JT-123-S 1:1 (600ohm) dual primaries in series - 9.43H dual secondaries in series - 9.43H Data provided by NY Dave
Conclusions With the exception of the Edcor and Mouser types (both of which are exceptionally cheap), all the transformers tested had sufficient primary inductance for their intended purpose. The Sowter and Carnhill examples were well specified and the BBC types even more so. Given their size and high values of inductance, the two Carnhill transformers tested represent excellent value for money. As a simple rule of thumb, for little loading at 20Hz, the inductance of a winding in Henries should be approximately its specified winding impedance divided by 100. So a 600 ohm winding should have an inductance of at least 6 Henries and a 10K winding should have an inductance of around 100 Henries. Ian Thompson-Bell January 2012 Version 0.2 5 th Feb 2012 added groupdiy members results