The SUT - MC Step Up Transformer - and how it works
Latest update 2 September 2020
Step-up transformers for MC cartridges (also called SUTs) are one of the most esoteric and often misunderstood hifi components. That might explain why they are so rarely used. (Several good active MC / RIAA amplifiers are definitely anopther explanation). Actually, it’s a shame, because a good transformer is often the optimal for a MC cartridge. Despite a lot of technical stuff in the following, the conclusions are quite easy to understand and the accompanying spreadsheet makes it easy for everyone to adapt the cartridge to the MC transformer and subsequently to the RIAA.
Talking about input and output impedances for a transformer are especially misunderstood. Look at a transformer like a gearbox. Without knowing what is connected to the input and output, one cannot calculate what is happening through the transformer. This applies to both amplification and impedances. A SUT quite simply is an impedance converter.
If you want to skip the explanations and calculate on a combination of cartridge and transformer, please go directly to the spreadsheet here: LINK SPREADSHEET
Why use a SUT?
The big disadvantage with MC cartridges is that they typically have 20 dB lower output than a MM cartridge. Necessarily we have to amplify the signal to suit the subsequent RIAA. This can be done actively with a so-called "headamp" (has nothing to do with headphones!) or with a SUT.
Many people believe that an optimally designed and adapted transformer is still the best solution. This is especially true when we talk about cartridges with very low output; 0.3 mV or lower. To this I can add. Yes, but it costs! The best transformers I have heard cost as much as the most expensive cartridges. The MYSL transformer seen on the left costs 41,200 DKR! (approx. 5,500 Euro).
Previously, it was almost impossible to achieve an optimal signal to noise ratio with an active circuit. A transformer was the only solution. A gain of 20 to 30 dB is not a problem but doing so with noisy tubes or transistors could be quite a challenge. Modern semiconductors can now much better live up to the requirements of low noise, but tubes should still preferably have a transformer at the input if you wish to use low output MC cartridges.
Active solutions will be treated another time, this is all about the SUT - the transformer solution.
In addition to the subject of noise, there is the actual sound quality that supporters of the SUTs empathies. The distortion in a transformer is of a completely different nature than what happens in an active amplifier. The harmonic distortion in a transformer is highest at low frequencies and drops rapidly upwards in frequency. Absolutely an advantage in a SUT where the RIAA-corrected signal is non-linear. An active circuit will typically behave quite the opposite way around with increasing distortion up in the higher frequencies. More importantly, the annoying intermodulation distortion is typically lower in a good transformer than in a transistor circuit. Although a transformer is not without distortion (nothing is), it is almost insignificant compared to what many transistor amplifiers can perform. What speaks against the use of transformers is sensitivity to radiation (hum) and the price! Transistors and ICs can cost up to a few bucks, but good transformers will always cost a whole lot more. Sometimes several 100 times more! This is due to the use of expensive core materials, as well as OCC copper or silver cables internally. If you buy a ready-made solution, it is often put in a nice box designed according to all the rules of art. Probably expensive, but really it does not have much to do with the sound.
Loading your cartridge
Before we look at the interface between the cartridge and the transformer, it is appropriate to take a look at the effect of different impedance loads on your MC cartridge.
When a electrical signal looks into an ohmic load, we get a voltage divider - the source impedance (also called the output impedance) looks into the load impedance (input impedance). The source here is a cartridge but in principle it can be other signal sources. This voltage divider acts as an attenuator, or a kind of fixed volume control. If the load impedance is much higher than the source impedance, we get a small attenuation. The golden rule in audio is to lead a signal into a load that is at least 10 times greater than the output impedance of the signal. This will avoid loss of signa. This also applies to our MC cartridges. If the ratio between the output and input impedance is 10 then we get an attenuation of approx. 1 dB. A loss we can accept, but on the other hand, if the source and load impedances are the same, we lose 6dB. This "rule" of 10 times the source impedance applies to active circuits. I have gradually come to doubt whether this also applies to transformers. For now, I will say no. Far higher as well as much lower impedances may work fine, but in the end - give it a try.
For a long time, the typical MC cartridge had an internal impedance (source impedance) around 10 ohms. But in recent years it has become much lower. Even the best cartridges from around the year 2000 can be down around the 4 ohms or lower (Accuphase and Transfiguration). If we talk about some of the latest top pickups, we find much lower impedances, all the way down to around 1 ohm (MYSL, LYRA and IKEDA). Certain exotic Japanese are now below 1 ohm.
Does the sound change when we change the impedance? Yes, definitely! Many people say that the sound is typically brighter and more shrill if the load impedance is too high. Conversely, a too low impedance will result in a dead and somewhat low energy reproduction with lack of air. Often, cartridge manufacturers will just state "higher than 50 ohms" or similar. Then it is up to the user to try it out. Again, I would say that there probably will be some differences between active circuits and SUTs.
The general rule of thumb that many manufacturers adhere to is an input impedance of 100 ohms. It is an appropriate value for most cartridges, and will very often be the impedance you find on typical MC-inputs. Personally I stick to the rule of 10-20 times the source impedance. However, it does not always come out as the best.
Many people are of the belief that the load should be smaller when using a transformer. Here I have seen 5 times 5 times the internal impedance and even down to 1 time - ie loaded with the same impedance as the cartridge. However, I would not recommend a load equal to (or close to) the output impedance of the cartridge. It drops the level with 6 dB and the sound becomes very flat and undynamic. Moreover, it is difficult to achieve with a transformer.
Transformer - winding and impedances
The turnover ratio in a transformer is the ratio between the primary and secondary windings. As mentioned in the beginning a kind of "electric gearbox". The voltage is transformed up or down in the same ratio as between the two windings. A transformer with e.g. 100 windings on the primary and 1,000 on the secondary side have a turnover ratio of 1:10. In other words, it will (under ideal conditions) transform the voltage up to 10 times as much as at the input. But a transformer is 100% passive and cannot draw energy out of thin air. The increased voltage will at the same time mean a corresponding reduction in the supplied current. Here we have what causes the impedance transformation in a transformer. The transformer itself does not have an impedance - it reflects the impedance from one side, so it is seen as another impedance on the other side. It works in both directions. In the above example with a 1:10 step-up transformer, with 47k load on the secondary side (your RIAA) will by the cartridge be seen as 470 ohms on the primary winding. In the same way, the output impedance will be larger. 10 ohms from your cartridge turns into an output impedance of 1000 ohms. As it looks into 47 kohm, this is no problem.
Examples of good SUTs:
IST-201 - ratio 1:20
Having read the above it is logical that a cartridge with an output signal of 0.5mV, when used with a SUT that converts with a ratio of 1:10, will give 5mV at the output. But no – this would take that the internal impedance of the cartridge is zero ohms. In practice, most MC cartridges are around 5 ohms and for transformers with a low turnover ratio (less than or equal to 1:20), we can calculate the turnover ratio of the transformer.
As mentioned, the internal impedance of the cartridge will never be zero. Therefore, this value must be included in the calculation if we want a completely correct result. Let's take an example:
The SUT has a 1:10 ratio and our cartridge says 10 ohms internal impedance. A normal load of 47k will then result in the cartridge seeing 470 ohms. The 470 ohms load and the 10 ohms source impedance forms a voltage divider. The level in this example decreases by only 0.2 dB (10 / (470 + 10) = 0.0208 times). It's not very much and not worth worrying about.
But if we have a transformer with a larger turnover ratio and a cartridge with a higher output impedance, then it can cause problems. Let's take a cartridge with 40 ohms internal impedance (eg an old Denon DL 103 from before 2007) and a transformer with 1:30 ratio (Intended for an Ortofon MC 3000). The 47k on the RIAA is then seen by the cartridge as 52 ohms. Therefore, the voltage drop due to mismatching will be 0.43 (approx. 5 dB). That's a lot - almost half the voltage the cartridge generates. Instead of the 30 times that the transformer's turnover ratio indicates, we only get approx. 13 times gain!
Although we have a transformer with a large turnover ratio, it does not fit the Denon DL-103 and will not provide the gain we expect. At the same time get a load that is close to the internal impedance. It may be the sound is OK, but we do not make optimal use of either cartridge or transformer.
Here we may have the reason for the misunderstanding that often arises. Many believe that a SUT needs to have a certain load, that will fit the cartridge. Yes, that's probably true if we're just looking to get the maximum signal out of the cartridge. However, we must not forget that we do not necessarily always go for the maximum signal. What we are aiming for is an appropriate signal for the subsequent RIAA as well as an optimal adjustment of the impedance to our pickup. Normally we will aim to get approx. 5 mV to the connected RIAA. Some can withstand more (typically tubes), but in the end it has to fit with the gain and which signal we want. Too much will clip the input - too little will make noise.
Let's try with the Denon DL-103 again, but this time with a Denon AU-1000 transformer (pictured here on the left) with a ratio of 1: 13.5. We then get a load of 258 ohms and an output signal to the RIAA of 3.51 mV (attenuated 1.25 dB due to the impedances). Here it looks somewhat more sensible and we still get a suitable output for our RIAA.
more optimal would be if we could change the input impedance of our RIAA to 100
kohm. You can do this in some RIAAs, and certainly if you build the amplifier
yourself. With a load of 100 kohm, our DL-103 through the AU-1000 will look
into 549 ohms and we get 3.8 mV at the input.
The big mistake you often see when choosing a transformer is that you focus too much at a single parameter. We must take into account both the gain obtained and the resulting impedance.
Transform with multiple primary and secondary windings
Above I have only mentioned SUTs with a fixed ratio. However many offer switches that will change impedance and gain. E.g. the Lundahl LL 1933 has 2 primary windings and 2 secondaries. It provides several options. You can connect the windings in series or parallel. Finally, it is also possible to use it for a fully balanced signal on both input and output. (Something I will come back to another time).
A very different example is my good old Fidelity Research FRT-3 transformer. Via a switch, 30 ohms or 16 ohms can be selected. What is this? Why doesn’t it say anything about ratio and gain? Well, it does, so to speak, but you just have to find it online:
ohms, 30 ohms and PASS (MM)
If you just put a turnover ratio of 1:36 into the calculations (see Excel spreadsheets) it does not fit at all. The impedance will then be 36 ohms. Can't they compute? Notice the "secondary 12 kohm". Internally in the transformer, there is a 16.2k resistor parallel to the output. It sits parallel above the 47 kohm of the following RIAA. Overall, it gives the mentioned 12k. With this new secondary load, the primary load is calculated at 9.3 ohms. Very close to the specified 10. In the second position 26 dB I get the load to just over 27 ohms - so again close to the specified 30.
Many especially older MC transformers or cartridges tended to oscillate at higher frequencies. One has to say this FR belongs to the slightly older ones. They liked to have the advantage of being loaded on the secondary side (lower impedances). This trick has been used by FR.
Further down the page, I get into impedance matching by adding resistors on the secondary side (and the primary side) of the SUT. It also appears on the spreadsheet.
Transformers aren’t perfect
The calculations above and in the attached spreadsheet are based on a "perfect" transformer. That is, with wire without resistance and with zero capacity in the windings, with no parasite inductance and infinite primary inductance and so on. In the real world, we have to accept the limitations of nature and work with wires with resistance, capacity, and so on. All transformers have limitations and those with very high turn ratios typically have more, as they contain several windings on the secondary side. This means greater resistance and capacity which will naturally limit the frequency range upwards.
Why not just fewer windings on the primary side? Because it will reduce the primary inductance and limit the frequency range downward. As a rule of thumb, a transformer with lower turn ratios will give better results both technically and acoustically.
Another phenomenon is magnetic hysteresis, which causes the magnetization of the core to come later than the electric field that creates the magnetism. One can say that there is an inertia in the magnetic material. It comes from the Greek word “hysteresis” (afterwards). All ferromagnetic materials used in transformer cores exhibit this phenomenon to a greater or lesser degree. The phenomenon is also known from tape recorders. This is one of the reasons why far more optimal core materials were developed - e.g. permalloy, cobalt nickel and others. This aspect is also made when making transformers.
That's why we talk about hysteresis loss in a transformer. By applying a magnetic field to the magnetic core in the transformer (via the primary winding), the core is magnetized. Starting from 0 - a non-magnetized core - the magnetism increases as the field is raised. When the field is lowered (the current turns) the magnetizing of the core should again drop to 0. Unfortunately, there are always minor impurities in the core which do not return to 0. Only when the opposite signal is sent to the primary winding we force the core to change polarity. Thus, every time we turn the current in our signal through the core, there will be some inertia, which causes distortion.
I don’t really know how much this really means for a SUT. Data for the best transformers does not mention the phenomenon at all. If you want to know more about this topic, search for "hysteresis" on the web.
Transformers and impedance matching
As I mentioned in the introduction, most SUTs are best used with a load of approx. 5-10 times their output impedance. However, this is an assumption, and lately I have had some doubts about whether this also applies to the use of SUTs. It is thus not certain that we will achieve this through the transformer, as we are definitely also striving to get the greatest possible gain. Now, however, it is the case that we (I) would like the opportunity to adapt the sound to the MC cartridge we use. It can be one with 5 ohms output impedance or one with 20 ohms. They may not require the same gain nor the same impedance.
As the case in the above-mentioned FRT-3, vu can then add a resistor parallel to the secondary output of the SUT (in parallel with the 47 kohms in our RIAA). This lowers the impedance. It is best if we build our RIAA ourselves and therefore can choose e.g. 22 kohm or 33 kohm input resistance. This will give us more options.
However, we must be aware that this cannot be used uncritically. As mentioned above, we must constantly make sure that we do not approach a load that corresponds to the output impedance of the cartridge. It will attenuate the signal by 6 dB and most likely give a very poor reproduction. Nor should we in this way try to adapt a transformer with high gain (higher turnover ratio) to a high-impedance cartridge with a high output level. In such cases, there is definitely a maladaptation. Adjustment with resistors may work, but keep in mind we are throwing gain away at a point in the signal path where we absolutely need gain. The optimum will always be to choose the right transformer for your cartridge. Also remember that it must be good resistors! When you only need 2, you can easily buy the best there is.
In some places I have seen recommendation of resistors across the primary winding. I have now introduced this option in the spreadsheet. It should be tried, but this could have the same disadvantages that loading of the secondary winding shows.
The link to the spreadsheet was in the top, but her you have it again - SUT SPREADSHEET
The link to the spreadsheet was in the top, but her you have it again - SUT SPREADSHEET